Annu. Rev. Astron. Astrophys. 1984. 22: 37-74 Copyright © 1984 by Annual Reviews. All rights reserved

4. STAR-FORMING HISTORIES

4.1 Constant Star Formation Rates

The localized bursts of star formation that characterize Irrs do not necessarily imply galaxy-wide bursts of activity; the global mean rates could still be constant while local variations are large (cf. 223). Gallagher et al. (126) have explored the star formation histories of a sample of high surface brightness Irrs chosen for their blue colors by considering parameters that measure the stellar birthrate over different time scales: The galaxian mass is a clue to the astration rates integrated over the galaxy's lifetime, the blue luminosity is dominated by stars formed over the last few billion years, and the ionizing photons give the current rate. They found that for most Irrs these parameters are consistent with a constant mean rate of star formation and IMF over the galaxy's lifetime, in agreement with results for the Magellanic Clouds (e.g. 283).

Few Irr systems seem to be undergoing honest global star formation bursts, i.e. only in unusual circumstances does the current global stellar production rate exceed the lifetime average rate (126, 191, 223, 283). In addition, Hunter (189) showed that the time scales to exhaust the present gas content at current astration rates and the metallicities were also consistent with constant stellar birthrates if all of the detected gas readily participates in the system's evolution. These results stand in contrast to studies based on colors and emission-line properties in which bursts and peculiar IMFs are found to be a normal feature of blue galaxies (see the following section). We are not currently able to reconcile these conclusions, although galaxy sample selections, data characteristics (i.e. local vs global measurements), and possible problems with stellar population models are all factors that may lead to differences.

If stars are formed at constant rates in Irrs, what does this imply about these galaxies? As we have seen, there is empirical evidence that gas densities must exceed some critical value for star formation to proceed at normal levels. A constant stellar birthrate then implies that a galaxy must maintain a constant amount of gas above this critical density, in a state suitable for starbirth, even as gas is continually being locked into new stars. How then does a galaxy manage to keep shuffling the same amount of gas into stars? We would expect that as more and more gas was turned into stars, the overall gas density would drop and the star formation rate would decrease. Thus the standard models predict that star formation should steeply decline with time in all galaxies (94, 126). This difficulty can be avoided either by postulating a continuous resupply of gas to star-forming regions, i.e. by maintaining constant average gas density (126, 221, 253), or by presuming the star formation rate is not solely dependent on mean gas density. Models of the former type require gas inflow from a halo or outer disk, which has never been confirmed by direct observation. In the latter class of models, the star formation rate must not depend on the total amount of gas. For example, the production of dense interstellar clouds from which stars are born could be the controlling factor and could be set by the OB stars themselves (e.g. 324).

A complementary question concerns the early stages of formation of an Irr. Given that the global star formation rates in the past might not have exceeded current levels by much, and that most Irrs probably formed as gravitationally bound entities at approximately the same time as other types of galaxies, then it seems that these systems managed to collapse into rotating ``puddles of gas'' before beginning serious star formation. Star formation in the earliest phases of galaxy evolution is still a heated issue with regard to all types of galaxies, and as Irrs represent the extreme end of the ``late-bloomers,'' they should provide an important point for empirical comparisons with galaxy formation models.